The realization of practical, optically controlled phased antenna arrays, which have received intensive interest due to their applications to microwave communication and radar systems, is currently hampered by the extreme complexity required in efficiently transmitting several hundred signals (or microwave delays) from the input (control) to the antenna array of the system. These difficulties are compounded by the demands of modern phased array systems including extremely high bandwidth broadcast frequencies (ranging from 20 to 60 GHz at up to 5 GHz bandwidth per channel), severe requirements on signal-to-noise and dynamic range, clutter cancellation, and the myriad requirements for beam forming including null steering, and generating multiple, squint-free beams at different frequencies. In addition, while problems regarding the transmission of the radar beam are relatively straightforward in their solution, there is still considerable controversy on the best means for receiving the return beam.
To date, several approaches have been suggested for solving optically controlled phased array antenna problems, although there have been only one or two practical demonstrations. Most architectures employ space division multiplexing to distribute a series of time or, less conveniently, phase delays to the antenna network; see, e.g., W. Ng et al, "The First Demonstration of an Optically Steered Microwave Phased Array Antenna Using True-Time-Delay", IEEE Journal of Lightwave Technology, Vol. 9, pp. 1124-1131 (1991) and C. Hemmi et al, "Optically Controlled Phased Array Beamforming Using Time Delay", Proceedings of DoD Fiber Optics Conference 1992, pp. 60-63, McLean, Va. (1992).
In these systems, the microwave delays are impressed on the optical carrier by using fibers or waveguide delay lines. This "true time delay" architecture then distributes the various delays to an antenna array via an optical switching matrix, such as a large integrated optic crossbar consisting of LiNbO.sub.3 switches or laser or detector arrays, or alternatively, liquid crystal spatial light modulator "stacks", followed by electronic amplifiers needed to compensate for the insertion losses (sometimes approaching 100 dB in large scale implementations) of the switch.
One additional difficulty with this space division multiplexing architecture is that all fibers which transport the microwave/optical signals from the input to antenna terminals must be of the same length so as not to introduce extraneous phase shifts, time delays, or noise into the signals. Hence, these architectures have been proven to be limited by losses, they are difficult to implement, they are bulky, and are extremely costly.
An alternative approach is based on wavelength division multiplexing; see, for example, H. Haus, "Proposed Scheme for Optically Controlled Phased Array Radar", 3rd Annual DARPA Symposium on Photonics Systems for Antenna Applications, Monterey, Calif., Jan. 20, 1993. This system employs pulses from a chirped, mode-locked laser. The beat frequency of two adjacent Fourier frequency components from the laser (e.g., the nth and (n+1)th components), is upshifted in phase by an amount .PHI..apprxeq.2n+1 from the beat frequency of the next two higher Fourier components. Thus, provided that one could fabricate tunable filters of sufficiently narrow spectral bandwidth to select a given Fourier component of the pulse spectrum, one could then extract all the phases at various antenna sites using pairs of tunable filters. These filters have been termed "channel dropping filters" (CDFs), and it was proposed by Haus, supra, that they be fabricated using semiconductor grating structures similar to those used in .lambda./4-shifted distributed feedback (DFB) lasers.
The difficulty with this approach is that it relies on very narrow bandwidth, tunable CDFs, which have yet to be demonstrated. One further problem is that this is a phase-delay architecture. True time delay is implementable, but only in a "coarse", wavelength multiplexed manner. These difficulties are counterbalanced, however, by the fact that all phase delays can be transmitted from input to antenna along a single fiber, and it is simple to utilize this concept in both transmit and receive modes.
Nevertheless, there remains a need for a phased array radar architecture which can accommodate hundreds of channels in a single fiber and process these hundreds of channels, which can substantially use existing components, and which can be configured in both a transmit and receive mode.